Plαntingα's Theory of Proper Names - Project Euclid

115 Notre Dame Journal of Formal Logic Volume 24, Number 1, January 1983

Plαntingα's Theory of Proper Names DAVID F. AUSTIN*

In "The Boethian compromise," Alvin Plantinga proposes a theory of proper names. His theory, he argues, is superior to theories of proper names suggested by the work of Mill, Donnellan, Kripke, and Kaplan in its handling of at least three puzzles: . . . those presented by empty (i.e., non-denoting) names, by negative existentials containing proper names, and by propositional identity in the context of propositional attitudes. ([7], p. 129) Plantinga also argues that his theory avoids one criticism which he takes to be very damaging to theories of proper names held by Russell and Frege. I will argue that Plantinga's theory is unsatisfactory in its handling of the puzzle presented by propositional identity in the context of propositional attitudes. In order to motivate Plantinga's theory, I will begin by giving a very brief statement of the criticism which Plantinga takes to be very damaging to Russell's and Frege's theories of naming. Then, I will state the puzzle as Plantinga renders it, presented by propositional identity in the context of propositional attitudes. Next, I will show how Plantinga's own theory avoids that criticism and at least appears to resolve the puzzle. Two objections to Plantinga's theory will then be presented. I will also consider some replies that Plantinga might reasonably make. In giving my objections, I have endeavored to present an "internal criticism" of Plantinga's view; that is, I have

*I want to thank Earl Conee and Herbert Heidelberger for several helpful discussions on the topic of this paper. Terence Parsons commented on an earlier draft, and I thank him for his comments. I am also grateful to the referee, who caught two errors of textual interpretation in the penultimate draft. My greatest debt of gratitude here is owed to Edmund Gettier, without whose encouragement and philosophical guidance this paper would not have been written.

Received November 1 7, 1980; revised November 2, 1981

116

DAVID F. AUSTIN

attempted to base my objections on assumptions that are either explicit or clearly implicit in Plantinga's presentation of his view. To simplify discussion, I will restrict this essay to cases involving nonempty terms (i.e., names and other terms that denote existing entities). 1 At the beginning of [7], Plantinga states an objection, much like one made by Kripke,1 which Plantinga says shows that " . . . no (definite) description of the sort Russell and Frege had in mind is semantically equivalent to a name like 'Socrates'" ([7], p. 129). The objection may be illustrated by means of the following argument. Suppose, for purposes of a reductio, that (1)

'Aristotle' is short for 'the teacher of Alexander the Great'. 2

Now it is contingently false that (2)

Aristotle is not the teacher of Alexander the Great.

But if (1) is true, then what (2) says is just that (3)

The teacher of Alexander the Great is not the teacher of Alexander the Great.

But (3) is necessarily false and (2) is not. So what (2) says is not just what (3) says. Hence, (1) is to be rejected. Similarly, the objection goes, for many other names and definite descriptions. (Of course, we need not suppose that the predicate is exactly what occurs in the description. Here, for example, 'is not a teacher' would have done as well as 'is not the teacher of Alexander the Great'. 3 ) According to Plantinga, then, one test that any satisfactory theory of proper names must pass is that it must avoid objections of the sort illustrated by (l)-(3). Let us call this test the "Necessity Test". 2 The other test of interest here is provided by a puzzle presented by propositional identity in the context of propositional attitudes. I will quote Plantinga's own rendering of the puzzle and will then reformulate it as an argument. Plantinga's rendering is as follows: If we think . . . that a proper name typically exhausts its semantic role in denoting its referent, then presumably the result of replacing it in a sentence like (8) Mark Twain was a pessimist or (9) Mark Twain is the same person as Samuel Clemens by another name of the same object will express the same proposition . . . But surely . . . (this) is wrong. Clearly a person could know the proposition expressed by (8) without knowing that expressed by (10) Samuel Langhorne Clemens was a pessimist . .. There are various expedients that might tempt anti-Fregeans here: none, I believe, is satisfactory. ([7], p. 131) Here is my reformulation. Let us suppose, first, that 'knows' is a two-place predicate ('£ knows p') that is true of ordered pairs of persons and proposi-

PLANTINGA'S THEORY OF PROPER NAMES

117

tions; and that the that-clauses below are terms denoting propositions expressed by the sentences embedded in the that-clauses. Assume, for reductio, that (4)

The only semantic contribution that the names 'Mark Twain' and 'Samuel Langhorne Clemens' make to the sentences 'Mark Twain was a pessimist' and 'Samuel Langhorne Clemens was a pessimist' is to denote their referents.

The other premises of the argument are: (5)

If the only semantic contribution that the names 'Mark Twain' and 'Samuel Langhorne Clemens' make to the sentences 'Mark Twain was a pessimist' and 'Samuel Langhorne Clemens was a pessimist' is to denote their referents, and Mark Twain = Samuel Langhorne Clemens, then the proposition that Mark Twain was a pessimist = the proposition that Samuel Langhorne Clemens was a pessimist.

(6)

(3S) 0 ((i) S knows that Mark Twain was a pessimist & (ii) ~S knows that Samuel Langhorne Clemens was a pessimist).

(7)

Mark Twain = Samuel Langhorne Clemens.

The negation of (4) is easily seen to follow. Hence, (4) is to be rejected. Similar arguments may be made for other pairs of codesignative names. Thus, according to Plantinga, a second test that any satisfactory theory of proper names must pass is that it must give a good explanation of supposed truths such as (6), and avoid assumptions such as (4). The linguistic behavior usually cited as evidence for (6), and which needs to be taken into account in developing any satisfactory theory of proper names, is of the following sort: Smith, a competent speaker of English who is ignorant of the fact that 'Mark Twain' and 'Samuel Langhorne Clemens' are two names for the same object, may sincerely and reflectively assent to 'Mark Twain was a pessimist' and yet dissent from or withhold assent to both 'Samuel Langhorne Clemens was a pessimist' and 'Mark Twain is the same person as Samuel Langhorne Clemens'. There are a number of ways to take such linguistic behavior into account. Among these ways are the various expedients to which Plantinga alludes in the passage quoted above. I believe that Plantinga himself was at one time tempted by one of these expedients. We may sketch it as follows: One might reject (6), while holding that (4), (5), and (7) are true (and while continuing to hold that 'knows' is a two-place predicate that is true of ordered pairs of persons and propositions) and attempt to explain the apparent truth of (6) in the following way. While no one could both know that Mark Twain was a pessimist and fail to know that Samuel Langhorne Clemens was a pessimist (because the propositions are one and the same), one might fail to know that the proposition expressed by 'Samuel Langhorne Clemens was a pessimist' is the proposition that Mark Twain was a pessimist, even though one knows the latter proposition. And one might fail to know this because one fails to know that Mark Twain was also named 'Samuel Langhorne Clemens' (though, presumably, one would not fail to know that Mark Twain is Samuel Langhorne Clemens, i.e., that Mark Twain is Mark Twain). Such reasoning

118

DAVID F. AUSTIN

about our knowledge of which sentences express which propositions would seem to provide one plausible way of explaining our inclination to assert, "Someone might know Mark Twain was a pessimist without knowing that Samuel Langhorne Clemens was a pessimist", without thereby committing ourselves to the truth of (6). Given, then, that the facts about linguistic behavior that must be explained are not necessarily those expressed by (6), let us redescribe this second test of a theory of proper names—call it the "Propositional Identity Test" (or, PIT)—as follows: there are facts about linguistic behavior that seem to be expressed by (6), and any satisfactory theory of proper names must take these facts into account; this can be done either by saying that (6) accurately describes these facts and is thus true; or, alternatively, by saying that (6) is false, but does not express the facts in question. Remarking on an explanation that takes the second alternative, Plantinga writes, "Now perhaps this is not wholly implausible; it does have about it, however, a certain air of the arcane. In any event a better explanation is available . . . " ([7], p. 135; see also [1]), an explanation, Plantinga indicates, that allows us to hold on to a 'simple truth' ([7], p. 135) like (6). The 'better explanation' is provided by Plantinga's own theory of proper names, and I now turn to a brief exposition of that theory (Section 3) to show how it appears to pass both the Necessity Test and the PIT (Section 4). 3 The relevant part of Plantinga's theory of proper names may be summarized as follows: (E)

Proper names express essences.

Plantinga's main proposal about which essences are expressed by proper names is: (A)

The essences that a proper name expresses are expressed by descriptions of the form 'the F-in-α', where 'F-in-α' is the α-transform of the predicate Ψ\

In what follows, it is the conjunction of (E) and (A) that I will mean by 'Plantinga's theory of proper names'.4 Plantinga does not offer a general account of which descriptions express the essences expressed by proper names, though he makes several suggestionsparasitic on views about proper names held by Russell, by Frege, by Searle, and by Donnellan and Kripke—about which descriptions one might associate with proper names. I will not discuss those suggestions here; it will suffice for my purposes in this essay to use syntactically simple suggestions parasitic on Russell's views. Both (A) and (E) (the α-transform principle and the essence principle, respectively) need some further explanation; specifically, more needs to be said about what Plantinga means by 'essence', 'express', and 'α-transform'. About what names and definite descriptions express, Plantinga has this to say: (8)

(a) A definite description, 'the F9 expresses the same property as does 'is the sole F\

PLANTINGA'S THEORY OF PROPER NAMES

119

(b) A proper name, N, expresses (in English) a property, F, if there is a definite description, D, in English or some extension of English, such that: (i) D expresses F and (ii) N and D are intersubstitutable salva propositione in sentences of the form Ί is F\ ([7], p. 134) (8b,ii) is based on the following principle of propositional identity, on which Plantinga seems clearly to rely: (9)

Proposition p = proposition q iff D (S)(A)((S is a person & A is a propositional attitude) D (S hasyl to p iff S has A to q)).

I will not try to say exactly how (8b, ii) is based on (9); I hope that the connection is clear enough for present purposes. Plantinga characterizes the notion of an essence of an existing entity as follows: (10)

e is an essence of x =^f (0 e is a property; (ii) O(x has e); (iii) Ώ(x exists D x has e); and (iv) Π(y)(y has e D x = y).

Plantinga characterizes notions of α-transform for both predicates and properties: (11)

The α-transform of a predicate, 'F', is 'F-in-α', where 'α' is a proper name of the actual world.

(12)

The α-transform of a property, F, is the world-indexed property being F-in-α, where ςα' is a proper name of the actual world.5

The following relationship is said to hold between α-transform predicates and properties: if the predicate ' F ' expresses the property F, then the α-transform of ' F ' (i.e., ςF-in-α') expresses the property being F-in-α (see [7], p. 133). Finally, (13)

x has property F in state of affairs w =